Draft 1

Loading Packages

library(tidyverse)
library(janitor)
library(readxl)
library(usmap)
library(stargazer)
library(haven)
library(tigris)
library(rio)
library(standardize)
library(effsize)
library(ggthemes)

Load and Clean Datasets

# Controls v2 from ACS
controls_2019 <- read_csv("controls_2019_v2.csv") %>%
  mutate(total_pop = SE_A00001_001) %>%
  mutate(percent_white = SE_A03001_002/total_pop) %>%
  mutate(percent_bachelors = SE_B12001_004/SE_B12001_001) %>%
  mutate(med_household_income = SE_A14006_001) %>%
  mutate(percent_hispanic = SE_B04001_010/total_pop) %>%
  select(Geo_FIPS, Geo_NAME, Geo_STATE, Geo_COUNTY, total_pop, 
         percent_white, percent_bachelors, med_household_income,
         percent_hispanic) %>%
  mutate(fips = as.numeric(Geo_FIPS))

cbp_2018 <- read_csv("cbp_2018.csv") %>%
  mutate(avg_wage = SE_T1200_001) %>%
  mutate(fips = as.numeric(Geo_FIPS)) %>%
  select(fips, avg_wage)

controls_2019 <- left_join(controls_2019, cbp_2018, by = "fips")

# Other Controls
poverty <- read_excel("PovertyEstimates.xls") %>%
  row_to_names(4) %>%
  clean_names() %>%
  mutate(fips = as.numeric(fip_stxt),
         poverty = as.numeric(pctpovall_2019)/100) %>%
  select(fips, poverty)

unemployment <- read_excel("Unemployment.xlsx") %>%
  row_to_names(4) %>%
  clean_names() %>%
  mutate(fips = as.numeric(fips_code),
         unemp = as.numeric(unemployment_rate_2019)/100) %>%
  select(fips, unemp)

controls_2019 <- controls_2019 %>%
  left_join(poverty, by = "fips") %>%
  left_join(unemployment, by = "fips")

# Import county migration data
c2c1519 <- import_list("c2c1519.xlsx")
c2c1216 <- import_list("c2c1216.xlsx")
c2c0812 <- import_list("c2c0812.xls")

# Rename columns
for (i in 1:52){
  colnames(c2c1519[[i]]) <- c("state_code_a", 
                                          "fips_a", 
                                          "state_code_b",
                                          "fips_b",
                                          "state_name_a",
                                          "county_name_a",
                                          "state_name_b",
                                          "county_name_b",
                                          "flow_ba",
                                          "flow_ba_moe",
                                          "flow_ab",
                                          "flow_ab_moe",
                                          "net_ba",
                                          "net_ba_moe",
                                          "gross_ab",
                                          "gross_ab_moe")
  colnames(c2c1216[[i]]) <- c("state_code_a", 
                                          "fips_a", 
                                          "state_code_b",
                                          "fips_b",
                                          "state_name_a",
                                          "county_name_a",
                                          "state_name_b",
                                          "county_name_b",
                                          "flow_ba",
                                          "flow_ba_moe",
                                          "flow_ab",
                                          "flow_ab_moe",
                                          "net_ba",
                                          "net_ba_moe",
                                          "gross_ab",
                                          "gross_ab_moe")
  colnames(c2c0812[[i]]) <- c("state_code_a", 
                                          "fips_a", 
                                          "state_code_b",
                                          "fips_b",
                                          "state_name_a",
                                          "county_name_a",
                                          "state_name_b",
                                          "county_name_b",
                                          "flow_ba",
                                          "flow_ba_moe",
                                          "flow_ab",
                                          "flow_ab_moe",
                                          "net_ba",
                                          "net_ba_moe",
                                          "gross_ab",
                                          "gross_ab_moe")
}

# Delete extraneous rows
for (i in 1:52){
  c2c1519[[i]] <- c2c1519[[i]] %>%
    filter(!row_number() %in% c(1, 2))
  c2c1216[[i]] <- c2c1216[[i]] %>%
    filter(!row_number() %in% c(1, 2))
  c2c0812[[i]] <- c2c0812[[i]] %>%
    filter(!row_number() %in% c(1, 2))
}

# Bind datasets together by year
c2c1519 <- bind_rows(c2c1519)
c2c1216 <- bind_rows(c2c1216)
c2c0812 <- bind_rows(c2c0812)

# Combine county migration data with controls and election data
df_2020 <- c2c1519 %>%
  filter(is.na(fips_b) == FALSE) %>%
  mutate(comb_fips_a = paste0(state_code_a, fips_a)) %>%
  group_by(county_name_a, comb_fips_a) %>%
  summarize(flow_ba = sum(as.numeric(flow_ba))) %>%
  arrange(comb_fips_a) %>%
  mutate(fips = as.numeric(comb_fips_a))

countypres2020 <- read_csv("countypres_2000-2020.csv") %>%
  filter(year == 2020) %>%
  filter(party == "DEMOCRAT" | party == "REPUBLICAN") %>%
  group_by(county_fips) %>%
  summarize(total_2p_votes = sum(candidatevotes)) %>%
  mutate(fips = as.numeric(county_fips)) %>%
  select(fips, total_2p_votes)

countypres2020_2p <- read_csv("countypres_2000-2020.csv") %>%
  filter(year == 2020) %>%
  filter(party == "DEMOCRAT") %>%
  mutate(fips = as.numeric(county_fips)) %>%
  group_by(fips) %>%
  summarize(candidatevotes = sum(candidatevotes)) %>%
  left_join(countypres2020, by = "fips") %>%
  mutate(dem_percent = candidatevotes/total_2p_votes)

countypres2020_2p_rep <- read_csv("countypres_2000-2020.csv") %>%
  filter(year == 2020) %>%
  filter(party == "REPUBLICAN") %>%
  mutate(fips = as.numeric(county_fips)) %>%
  group_by(fips) %>%
  summarize(candidatevotes = sum(candidatevotes)) %>%
  left_join(countypres2020, by = "fips") %>%
  mutate(rep_percent = candidatevotes/total_2p_votes) %>%
  select(fips, rep_percent)
  

df_2020_election <- left_join(df_2020, countypres2020_2p, by = "fips")

df_2020_combined <- left_join(df_2020_election, controls_2019, by = "fips") %>%
  mutate(avg_immi_rate = flow_ba/total_pop)

df <- df_2020_combined %>%
  filter(Geo_STATE != "02" && Geo_STATE != "72")

write.csv(df, "migration_df.csv")

Immigration

# Basic linear regression
stargazer(lm(dem_percent ~ avg_immi_rate, data = df_2020_combined, weight = total_pop),
          type = "text")

===============================================
                        Dependent variable:    
                    ---------------------------
                            dem_percent        
-----------------------------------------------
avg_immi_rate                -1.542***         
                              (0.132)          
                                               
Constant                     0.609***          
                              (0.008)          
                                               
-----------------------------------------------
Observations                   3,113           
R2                             0.042           
Adjusted R2                    0.042           
Residual Std. Error     53.990 (df = 3111)     
F Statistic          136.485*** (df = 1; 3111) 
===============================================
Note:               *p<0.1; **p<0.05; ***p<0.01

# Visualization of relationship between immigration rate and dem percent
ggplot(df_2020_combined, aes(x = avg_immi_rate, y = dem_percent, size = total_pop)) +
  geom_point(alpha = .3) +
  geom_smooth(method = "lm", aes(weight = total_pop), show.legend = FALSE) +
  labs(title = "Fall in Democratic Voting with Increase in Rate of Immigration",
       x = "Immigration Rate",
       y = "Two-Party Democratic Vote Share",
       subtitle = "2020 Presidential Election",
       caption = "Weighted by County Population")

Check for missing value in the economic data Check Alaska!! What are the patterns we expect to see if this is true, does the data look like what it would if this is happening Have a hypothesis that this should impact behavior, do you see a pattern that isn’t explained by other things Matching Viriginia reports cities and counties around them differently and doesn’t line up - drop ones that don’t align cleanly (fips codes matched things are weird, census has the full county but the data is reported split between the city and everything else) Know things about the places that people are coming from, -1, 0, 1 net partisanship of the people moving into an area, sorting, impact on people there

# Regression with controls
stargazer(lm(dem_percent ~ avg_immi_rate + percent_white + percent_hispanic +
               percent_bachelors + avg_wage + poverty + unemp + med_household_income, data = df_2020_combined,
             weight = total_pop),
          lm(dem_percent ~ avg_immi_rate + percent_white + percent_hispanic +
               percent_bachelors, data = df_2020_combined,
             weight = total_pop),
          lm(dem_percent ~ avg_immi_rate + avg_wage + poverty + unemp + med_household_income, data = df_2020_combined,
             weight = total_pop),
          type = "text")

======================================================================================================
                                                    Dependent variable:                               
                     ---------------------------------------------------------------------------------
                                                        dem_percent                                   
                                 (1)                         (2)                        (3)           
------------------------------------------------------------------------------------------------------
avg_immi_rate                 -0.663***                   -0.668***                  -0.589***        
                               (0.072)                     (0.074)                    (0.107)         
                                                                                                      
percent_white                 -0.476***                   -0.519***                                   
                               (0.011)                     (0.010)                                    
                                                                                                      
percent_hispanic              0.171***                    0.173***                                    
                               (0.010)                     (0.010)                                    
                                                                                                      
percent_bachelors             1.098***                    0.796***                                    
                               (0.024)                     (0.014)                                    
                                                                                                      
avg_wage                     0.00000***                                             0.00001***        
                              (0.00000)                                              (0.00000)        
                                                                                                      
poverty                       -0.230***                                              1.547***         
                               (0.066)                                                (0.093)         
                                                                                                      
unemp                         1.742***                                                 0.004          
                               (0.159)                                                (0.244)         
                                                                                                      
med_household_income         -0.00000***                                            0.00000***        
                              (0.00000)                                              (0.00000)        
                                                                                                      
Constant                      0.618***                    0.652***                   -0.155***        
                               (0.022)                     (0.011)                    (0.025)         
                                                                                                      
------------------------------------------------------------------------------------------------------
Observations                    3,112                       3,113                      3,112          
R2                              0.778                       0.745                      0.440          
Adjusted R2                     0.778                       0.744                      0.439          
Residual Std. Error      26.005 (df = 3103)          27.879 (df = 3108)         41.318 (df = 3106)    
F Statistic          1,361.864*** (df = 8; 3103) 2,267.819*** (df = 4; 3108) 487.790*** (df = 5; 3106)
======================================================================================================
Note:                                                                      *p<0.1; **p<0.05; ***p<0.01

# Regression standardized
df_2020_combined$dem_percent_scaled <- scale(df_2020_combined$dem_percent)[, 1]
df_2020_combined$avg_immi_rate_scaled <- scale(df_2020_combined$avg_immi_rate)[, 1]
df_2020_combined$percent_white_scaled <- scale(df_2020_combined$percent_white)[, 1]
df_2020_combined$percent_bachelors_scaled <- scale(df_2020_combined$percent_bachelors)[, 1]
df_2020_combined$med_household_income_scaled <- scale(df_2020_combined$med_household_income)[, 1]
df_2020_combined$percent_hispanic_scaled <- scale(df_2020_combined$percent_hispanic)[, 1]
df_2020_combined$avg_wage_scaled <- scale(df_2020_combined$avg_wage)[, 1]
df_2020_combined$poverty_scaled <- scale(df_2020_combined$poverty)[, 1]
df_2020_combined$unemp_scaled <- scale(df_2020_combined$unemp)[, 1]

stargazer(lm(dem_percent_scaled ~ avg_immi_rate_scaled + percent_white_scaled + percent_bachelors_scaled + 
               med_household_income_scaled + percent_hispanic_scaled +
               avg_wage_scaled + poverty_scaled + unemp_scaled, data = df_2020_combined,
          weight = total_pop), type = "text")

=======================================================
                                Dependent variable:    
                            ---------------------------
                                dem_percent_scaled     
-------------------------------------------------------
avg_immi_rate_scaled                 -0.116***         
                                      (0.013)          
                                                       
percent_white_scaled                 -0.504***         
                                      (0.012)          
                                                       
percent_bachelors_scaled             0.642***          
                                      (0.014)          
                                                       
med_household_income_scaled          -0.257***         
                                      (0.018)          
                                                       
percent_hispanic_scaled              0.204***          
                                      (0.011)          
                                                       
avg_wage_scaled                      0.068***          
                                      (0.010)          
                                                       
poverty_scaled                       -0.082***         
                                      (0.023)          
                                                       
unemp_scaled                         0.192***          
                                      (0.018)          
                                                       
Constant                             0.239***          
                                      (0.013)          
                                                       
-------------------------------------------------------
Observations                           3,112           
R2                                     0.778           
Adjusted R2                            0.778           
Residual Std. Error             160.019 (df = 3103)    
F Statistic                 1,361.864*** (df = 8; 3103)
=======================================================
Note:                       *p<0.1; **p<0.05; ***p<0.01

Emigration

df_2020_emi <- c2c1519 %>%
  filter(is.na(fips_b) == FALSE) %>%
  mutate(comb_fips_a = paste0(state_code_a, fips_a)) %>%
  group_by(county_name_a, comb_fips_a) %>%
  summarize(flow_ab = sum(as.numeric(flow_ab))) %>%
  arrange(comb_fips_a) %>%
  mutate(fips = as.numeric(comb_fips_a))

`summarise()` has grouped output by 'county_name_a'. You can override using the `.groups` argument.

df_2020_election_emi <- left_join(df_2020_emi, countypres2020_2p, by = "fips")

df_2020_combined_emi <- left_join(df_2020_election_emi, controls_2019, by = "fips") %>%
  mutate(avg_emi_rate = flow_ab/total_pop)

df_emi <- df_2020_combined_emi %>%
  filter(Geo_STATE != "02") %>%
  filter(Geo_STATE != "72")
write.csv(df_emi, "migration_df2.csv")

# Basic regression
stargazer(lm(dem_percent ~ avg_emi_rate, data = df_2020_combined_emi, weight = total_pop), type = "text")

===============================================
                        Dependent variable:    
                    ---------------------------
                            dem_percent        
-----------------------------------------------
avg_emi_rate                 -0.448**          
                              (0.178)          
                                               
Constant                     0.549***          
                              (0.010)          
                                               
-----------------------------------------------
Observations                   3,113           
R2                             0.002           
Adjusted R2                    0.002           
Residual Std. Error     55.106 (df = 3111)     
F Statistic           6.298** (df = 1; 3111)   
===============================================
Note:               *p<0.1; **p<0.05; ***p<0.01

ggplot(df_2020_combined_emi, aes(x = avg_emi_rate, y = dem_percent, size = total_pop)) +
  geom_point(alpha = .3) +
  geom_smooth(method = "lm", aes(weight = total_pop), show.legend = FALSE) +
  labs(title = "Fall in Democratic Voting with Increase in Rate of Emigration",
       x = "Emigration Rate",
       y = "Two-Party Democratic Vote Share",
       subtitle = "2020 Presidential Election",
       caption = "Weighted by County Population")

stargazer(lm(dem_percent ~ avg_emi_rate + percent_white + percent_hispanic +
               percent_bachelors + avg_wage + poverty + unemp + med_household_income, data = df_2020_combined_emi,
             weight = total_pop), 
          lm(dem_percent ~ avg_emi_rate + percent_white + percent_hispanic +
               percent_bachelors, data = df_2020_combined_emi,
             weight = total_pop),
          lm(dem_percent ~ avg_emi_rate + avg_wage + poverty + unemp + med_household_income, data = df_2020_combined_emi,
             weight = total_pop),
          type = "text")

======================================================================================================
                                                    Dependent variable:                               
                     ---------------------------------------------------------------------------------
                                                        dem_percent                                   
                                 (1)                         (2)                        (3)           
------------------------------------------------------------------------------------------------------
avg_emi_rate                  -0.538***                   -0.591***                  -0.417***        
                               (0.094)                     (0.098)                    (0.140)         
                                                                                                      
percent_white                 -0.485***                   -0.533***                                   
                               (0.011)                     (0.010)                                    
                                                                                                      
percent_hispanic              0.176***                    0.180***                                    
                               (0.010)                     (0.010)                                    
                                                                                                      
percent_bachelors             1.073***                    0.796***                                    
                               (0.024)                     (0.015)                                    
                                                                                                      
avg_wage                     0.00000***                                             0.00001***        
                              (0.00000)                                              (0.00000)        
                                                                                                      
poverty                       -0.237***                                              1.558***         
                               (0.066)                                                (0.093)         
                                                                                                      
unemp                         1.826***                                                 0.167          
                               (0.161)                                                (0.245)         
                                                                                                      
med_household_income         -0.00000***                                            0.00000***        
                              (0.00000)                                              (0.00000)        
                                                                                                      
Constant                      0.607***                    0.656***                   -0.183***        
                               (0.023)                     (0.012)                    (0.025)         
                                                                                                      
------------------------------------------------------------------------------------------------------
Observations                    3,112                       3,113                      3,112          
R2                              0.775                       0.741                      0.436          
Adjusted R2                     0.774                       0.741                      0.435          
Residual Std. Error      26.217 (df = 3103)          28.081 (df = 3108)         41.457 (df = 3106)    
F Statistic          1,333.636*** (df = 8; 3103) 2,224.033*** (df = 4; 3108) 480.322*** (df = 5; 3106)
======================================================================================================
Note:                                                                      *p<0.1; **p<0.05; ***p<0.01

Maps

plot_usmap(data = df_2020_combined, values = "avg_immi_rate", size = .1) +
  scale_fill_continuous(low = "white", "high" = "black") +
  labs(title = "Internal Immigration Rates from 2015-2019")

plot_usmap(data = df_2020_combined_emi, values = "avg_emi_rate", size = .1) +
  scale_fill_continuous(low = "white", "high" = "black") +
  labs(title = "Internal Emigration Rates from 2015-2019")

Regressing on Controls

stargazer(lm(percent_white ~ avg_emi_rate, data = df_2020_combined_emi, weight = total_pop),
          lm(percent_hispanic ~ avg_emi_rate, data = df_2020_combined_emi, weight = total_pop),
          lm(percent_bachelors ~ avg_emi_rate, data = df_2020_combined_emi, weight = total_pop),
          type = "text")

the condition has length > 1 and only the first element will be usedthe condition has length > 1 and only the first element will be usedlength of NULL cannot be changedlength of NULL cannot be changedlength of NULL cannot be changedlength of NULL cannot be changedlength of NULL cannot be changed

================================================================================
                                              Dependent variable:               
                                ------------------------------------------------
                                percent_white percent_hispanic percent_bachelors
                                     (1)            (2)               (3)       
--------------------------------------------------------------------------------
avg_emi_rate                      0.478***       -3.221***         1.322***     
                                   (0.170)        (0.184)           (0.113)     
                                                                                
Constant                          0.698***        0.366***         0.248***     
                                   (0.010)        (0.011)           (0.006)     
                                                                                
--------------------------------------------------------------------------------
Observations                        3,220          3,220             3,220      
R2                                  0.002          0.087             0.041      
Adjusted R2                         0.002          0.086             0.041      
Residual Std. Error (df = 3218)    53.085          57.426           35.105      
F Statistic (df = 1; 3218)        7.883***       305.654***       137.843***    
================================================================================
Note:                                                *p<0.1; **p<0.05; ***p<0.01

stargazer(lm(percent_white ~ avg_emi_rate, data = df_2020_combined_emi, weight = total_pop),
          lm(percent_hispanic ~ avg_emi_rate, data = df_2020_combined_emi, weight = total_pop),
          lm(percent_bachelors ~ avg_emi_rate, data = df_2020_combined_emi, weight = total_pop),
          type = "text")

the condition has length > 1 and only the first element will be usedthe condition has length > 1 and only the first element will be usedlength of NULL cannot be changedlength of NULL cannot be changedlength of NULL cannot be changedlength of NULL cannot be changedlength of NULL cannot be changed

================================================================================
                                              Dependent variable:               
                                ------------------------------------------------
                                percent_white percent_hispanic percent_bachelors
                                     (1)            (2)               (3)       
--------------------------------------------------------------------------------
avg_emi_rate                      0.478***       -3.221***         1.322***     
                                   (0.170)        (0.184)           (0.113)     
                                                                                
Constant                          0.698***        0.366***         0.248***     
                                   (0.010)        (0.011)           (0.006)     
                                                                                
--------------------------------------------------------------------------------
Observations                        3,220          3,220             3,220      
R2                                  0.002          0.087             0.041      
Adjusted R2                         0.002          0.086             0.041      
Residual Std. Error (df = 3218)    53.085          57.426           35.105      
F Statistic (df = 1; 3218)        7.883***       305.654***       137.843***    
================================================================================
Note:                                                *p<0.1; **p<0.05; ***p<0.01

stargazer(lm(avg_wage ~ avg_emi_rate, data = df_2020_combined_emi, weight = total_pop),
          lm(poverty ~ avg_emi_rate, data = df_2020_combined_emi, weight = total_pop),
          lm(unemp ~ avg_emi_rate, data = df_2020_combined_emi, weight = total_pop),
          lm(med_household_income ~ avg_emi_rate, data = df_2020_combined_emi, weight = total_pop),
          type = "text")

the condition has length > 1 and only the first element will be usedthe condition has length > 1 and only the first element will be usedlength of NULL cannot be changedlength of NULL cannot be changedlength of NULL cannot be changedlength of NULL cannot be changedlength of NULL cannot be changed

==========================================================================================================================
                                                             Dependent variable:                                          
                    ------------------------------------------------------------------------------------------------------
                            avg_wage                  poverty                    unemp             med_household_income   
                               (1)                      (2)                       (3)                       (4)           
--------------------------------------------------------------------------------------------------------------------------
avg_emi_rate               15,384.420                -0.171***                 -0.172***               73,601.550***      
                          (15,432.640)                (0.047)                   (0.013)                (18,886.240)       
                                                                                                                          
Constant                  49,381.010***               0.133***                 0.047***                61,268.500***      
                            (889.940)                 (0.003)                   (0.001)                 (1,089.132)       
                                                                                                                          
--------------------------------------------------------------------------------------------------------------------------
Observations                  3,218                    3,141                     3,219                     3,220          
R2                           0.0003                    0.004                     0.052                     0.005          
Adjusted R2                 -0.00000                   0.004                     0.052                     0.004          
Residual Std. Error 4,809,552.000 (df = 3216)    14.755 (df = 3139)        4.036 (df = 3217)     5,887,649.000 (df = 3218)
F Statistic           0.994 (df = 1; 3216)    13.012*** (df = 1; 3139) 176.189*** (df = 1; 3217) 15.187*** (df = 1; 3218) 
==========================================================================================================================
Note:                                                                                          *p<0.1; **p<0.05; ***p<0.01

Thoughts - add cost of living interacting with avg_wage?

stargazer(lm(avg_immi_rate ~ avg_wage, data = df_2020_combined, weight = total_pop),
          lm(avg_immi_rate ~ poverty, data = df_2020_combined, weight = total_pop),
          lm(avg_immi_rate ~ unemp, data = df_2020_combined, weight = total_pop),
          lm(avg_immi_rate ~ med_household_income, data = df_2020_combined, weight = total_pop),
          type = "text")

the condition has length > 1 and only the first element will be usedthe condition has length > 1 and only the first element will be usedlength of NULL cannot be changedlength of NULL cannot be changedlength of NULL cannot be changedlength of NULL cannot be changedlength of NULL cannot be changed

=====================================================================================================================
                                                           Dependent variable:                                       
                     ------------------------------------------------------------------------------------------------
                                                              avg_immi_rate                                          
                               (1)                      (2)                       (3)                    (4)         
---------------------------------------------------------------------------------------------------------------------
avg_wage                   -0.00000***                                                                               
                            (0.00000)                                                                                
                                                                                                                     
poverty                                              -0.032***                                                       
                                                      (0.009)                                                        
                                                                                                                     
unemp                                                                          -0.432***                             
                                                                                (0.030)                              
                                                                                                                     
med_household_income                                                                                    -0.000       
                                                                                                      (0.00000)      
                                                                                                                     
Constant                     0.066***                 0.059***                 0.071***                0.055***      
                             (0.001)                  (0.001)                   (0.001)                (0.001)       
                                                                                                                     
---------------------------------------------------------------------------------------------------------------------
Observations                  3,218                    3,141                     3,219                  3,220        
R2                            0.023                    0.004                     0.060                 0.00001       
Adjusted R2                   0.022                    0.004                     0.060                 -0.0003       
Residual Std. Error     7.204 (df = 3216)        7.305 (df = 3139)         7.062 (df = 3217)      7.285 (df = 3218)  
F Statistic          74.375*** (df = 1; 3216) 13.015*** (df = 1; 3139) 207.060*** (df = 1; 3217) 0.040 (df = 1; 3218)
=====================================================================================================================
Note:                                                                                     *p<0.1; **p<0.05; ***p<0.01

ggplot(df_2020_combined, aes(x = avg_immi_rate, y = percent_white, size = total_pop)) +
  geom_point(alpha = .3) +
  geom_smooth(method = "lm", aes(weight = total_pop), show.legend = FALSE) +
  labs(title = "Immigration Rate is Higher in Counties with Higher White Proportion",
       x = "Immigration Rate",
       y = "Proportion White",
       subtitle = "2019",
       caption = "Weighted by County Population")

ggplot(df_2020_combined, aes(x = avg_immi_rate, y = percent_hispanic, size = total_pop)) +
  geom_point(alpha = .3) +
  geom_smooth(method = "lm", aes(weight = total_pop), show.legend = FALSE) +
  labs(title = "Immigration Rate is Lower in Counties with Higher Hispanic Proportion",
       x = "Immigration Rate",
       y = "Proportion Hispanic",
       subtitle = "2019",
       caption = "Weighted by County Population")

ggplot(df_2020_combined_emi, aes(x = avg_emi_rate, y = percent_white, size = total_pop)) +
  geom_point(alpha = .3) +
  geom_smooth(method = "lm", aes(weight = total_pop), show.legend = FALSE) +
  labs(title = "Emigration Rate is Higher in Counties with Higher White Proportion",
       x = "Emigration Rate",
       y = "Proportion White",
       subtitle = "2019",
       caption = "Weighted by County Population")

ggplot(df_2020_combined_emi, aes(x = avg_emi_rate, y = percent_hispanic, size = total_pop)) +
  geom_point(alpha = .3) +
  geom_smooth(method = "lm", aes(weight = total_pop), show.legend = FALSE) +
  labs(title = "Emigration Rate is Lower in Counties with Higher Hispanic Proportion",
       x = "Emigration Rate",
       y = "Proportion Hispanic",
       subtitle = "2019",
       caption = "Weighted by County Population")

ggplot(df_2020_combined, aes(x = avg_immi_rate, y = percent_bachelors, size = total_pop)) +
  geom_point(alpha = .3) +
  geom_smooth(method = "lm", aes(weight = total_pop), show.legend = FALSE) +
  labs(title = "Immigration Rate is Higher in Counties with Higher Bachelor's Proportion",
       x = "Immigration Rate",
       y = "Percent Bachelor's Degree",
       subtitle = "2019",
       caption = "Weighted by County Population")

ggplot(df_2020_combined_emi, aes(x = avg_emi_rate, y = percent_bachelors, size = total_pop)) +
  geom_point(alpha = .3) +
  geom_smooth(method = "lm", aes(weight = total_pop), show.legend = FALSE) +
  labs(title = "Emigration Rate is Higher in Counties with Higher Bachelor's Proportion",
       x = "Emigration Rate",
       y = "Percent Bachelor's Degree",
       subtitle = "2019",
       caption = "Weighted by County Population")